Duke University Religion Index (DUREL)

1. Abstract

The Duke University Religion Index (DUREL) is a brief, five-item psychometric screening instrument developed by Harold G. Koenig and colleagues to capture the multidimensional nature of religiosity in epidemiological, public health, and clinical settings. Recognizing that single-item measures fail to distinguish public communal behaviors from private devotion and underlying motivational dispositions, the DUREL quantifies three distinct dimensions of religious involvement: Organizational Religious Activity (ORA; Item 1), Non-Organizational Religious Activity (NORA; Item 2), and Intrinsic Religiosity (IR; Items 3, 4, and 5). Items 1 and 2 utilize 6-point ordinal frequency scales ranging from 1 to 6, whereas Items 3 through 5 utilize a 5-point Likert agreement scale ranging from 1 (Definitely not true) to 5 (Definitely true of me).

Psychometric evaluations across numerous international samples demonstrate exceptional reliability and structural validity. The three-item Intrinsic Religiosity subscale regularly demonstrates internal consistency with Cronbach’s alpha coefficients between .75 and .91, as well as high two-week test-retest reliability (r = .91). Confirmatory factor analyses (CFA) consistently validate the three-factor oblique structural model across distinct sociocultural, linguistic, and religious cohorts, including Christian, Muslim, and secular populations. Construct and criterion validities are evidenced by consistent inverse correlations with clinical depression, state and trait anxiety, cognitive impairment, and all-cause mortality, alongside positive associations with physical health outcomes, subjective well-being, and social support. Designed to minimize respondent burden, the DUREL requires less than two minutes to complete, establishing it as one of the most widely employed psychometric inventories in psychiatric epidemiology, behavioral medicine, and the sociology of religion.

2. Keywords

Duke University Religion Index, DUREL, Harold G. Koenig, religiosity measurement, intrinsic religiosity, organizational religious activity, non-organizational religious activity, psychiatric epidemiology, behavioral medicine, psychometrics

3. Authors

The primary architect of the Duke University Religion Index is Harold G. Koenig, M.D., M.H.Sc. Dr. Koenig is Professor of Psychiatry and Behavioral Sciences and Associate Professor of Medicine at Duke University Medical Center in Durham, North Carolina. He serves as the Director of the Center for Spirituality, Theology and Health at Duke University. Dr. Koenig is globally acknowledged as a foundational pioneer in the empirical study of the nexus between religion, spirituality, and mental/physical health, having authored over 500 peer-reviewed scientific articles and dozens of academic volumes on the subject. Correspondence regarding the instrument has historically been directed to Duke University Medical Center (E-mail: [email protected]).

Co-investigators and international collaborators have played an instrumental role in refining, testing, and adapting the psychometric properties of the instrument across diverse cross-cultural and clinical milieus. Notable academic partners include Arndt Büssing, M.D., Professor of Quality of Life, Spirituality and Coping at Witten/Herdecke University in Germany, who co-authored pivotal benchmark papers analyzing the epidemiological utility and structural validation of the index. In addition, global validation studies have been led by international research teams, such as Amir H. Pakpour, Mohsen Saffari, and Sholeh Hafizi, who confirmed the instrument’s cross-cultural stability and measurement invariance across non-Western and Muslim-majority populations.

4. Purpose

The core purpose of the Duke University Religion Index is to provide researchers, sociologists, behavioral scientists, and healthcare clinicians with a brief, empirically robust, multidimensional measurement tool capable of assessing religious involvement within large-scale demographic surveys and clinical epidemiological studies. For decades, biomedical research was hindered by methodological oversimplification, frequently reducing complex religious phenomena to a single dichotomous item (e.g., religious denomination or “yes/no” religious affiliation). Such reductive classifications inevitably obscured the profound behavioral, social, and psychological variance that exists within religious traditions and failed to isolate which exact facets of religious life impact physiological resilience, psychiatric vulnerability, and clinical prognoses.

To overcome these limitations without imposing the severe administrative burdens associated with lengthy psychometric batteries—such as the 38-item Religious Status Inventory or the 20-item Religious Orientation Scale—Koenig and his colleagues designed the DUREL. The primary objective was to engineer an instrument that could be administered in under two minutes, either as a self-report paper-and-pencil questionnaire, within an electronic clinical intake portal, or via structured telephone interviews. Its brief five-item architecture was explicitly formulated to eliminate survey fatigue among frail, elderly, or severely ill medical inpatients, who represent primary cohorts in epidemiological longevity and health services research.

In clinical practice and biomedical research, the DUREL serves several vital purposes:

• Epidemiological Risk Stratification: Disentangling public, institutionalized behaviors from internal cognitive-motivational commitments allows epidemiologists to evaluate the individual predictive values of social integration versus personal coping mechanics on biomedical endpoints such as cardiovascular disease, immune functioning, systemic inflammation (e.g., C-reactive protein, interleukin-6), and survival rates in oncological contexts.
• Psychiatric Assessment: Facilitating the precise identification of patients who rely heavily on religious coping, thereby enabling psychiatrists and psychologists to understand intrinsic motivational frameworks, assess moral injury, address spiritual struggles, and leverage faith resources during cognitive behavioral therapy (CBT).
• Sociological and Cross-Sectional Mapping: Permitting cross-cultural comparisons regarding how institutional secularization affects public worship while private prayer or intrinsic commitments remain stable or follow independent trajectories.

5. Psychological Construct

The psychological construct operationalized by the DUREL is religiosity, conceptualized not as a monolithic identity, but as a complex multidimensional behavioral and cognitive phenomenon. Religious commitment manifests through disparate behavioral channels, cognitive schema, and affective states. The DUREL systematically separates this overarching construct into three psychometrically distinct and empirically non-redundant dimensions:

1. Organizational Religious Activity (ORA)

Operationalized through Item 1 (“How often do you attend church, synagogue, or other religious meetings?”), this dimension quantifies participation in public, communal religious ceremonies organized within formal, institutionalized structures. Organizational religiosity captures the extent to which an individual participates in collective social worship. Psychologically and sociologically, ORA serves as an indicator of an individual’s embeddedness within a faith-based collective. High ORA scores reflect active engagement in communal rituals, yielding robust benefits related to social capital, perceived social support, informal mutual aid, peer socialization, and shared communal norms that discourage high-risk health behaviors (such as substance abuse, smoking, or reckless behavior).

2. Non-Organizational Religious Activity (NORA)

Operationalized through Item 2 (“How often do you spend time in private religious activities, such as prayer, meditation, or Bible study?”), this subscale assesses the frequency of individual, self-initiated spiritual behaviors performed privately outside public institutional contexts. NORA reflects an autonomous, internalized devotion that does not rely on collective participation. This dimension is especially critical when examining populations who are functionally homebound, hospitalized, or geographically isolated, where physical attendance at religious institutions (ORA) becomes physically prohibitive. High NORA reveals private spiritual discipline, introspective meditation, personal devotion, and internalized ritualistic coping mechanisms utilized during instances of acute psychological distress or isolation.

3. Intrinsic Religiosity (IR)

Operationalized through Items 3, 4, and 5, this dimension assesses the internalization of religious faith as the central organizing motivation of the individual’s personality, life philosophy, and daily conduct. Rooted conceptually in the seminal psychological works of Gordon Allport and J. Michael Ross, intrinsic religiosity measures the extent to which religion is lived as an ultimate end in itself, rather than utilized as an instrumental means toward an external or utilitarian goal (such as social status, business networking, or familial compliance). The three items capture distinct cognitive and experiential facets:

• Item 3 (“In my life, I experience the presence of the Divine [i.e., God]”) captures the subjective-mystical and experiential realm of religious feeling, assessing perceived relational closeness to a transcendent entity.
• Item 4 (“My religious beliefs are what really lie behind my whole approach to life”) assesses the cognitive schema dimension, evaluating whether core doctrinal frameworks dictate personal worldviews, interpretive cognitive appraisals, and ultimate existential meaning.
• Item 5 (“I try hard to carry my religion over into all other aspects of my life”) evaluates operational integration, gauging the respondent’s intentional effort to apply religious ethical standards, behavioral boundaries, and virtues across social, relational, and vocational domains.

6. Theoretical Framework

The theoretical architecture of the DUREL is grounded in the classic psychology of religion, most notably Gordon Allport‘s religious orientation theory (Allport & Ross, 1967). Allport posited that individuals orient themselves toward religion in fundamentally diverging manners: extrinsic religiosity, wherein an individual “uses” religion for utilitarian, socio-emotional, or self-serving ends (such as obtaining community approval, comfort, or social standing); and intrinsic religiosity, wherein an individual “lives” their faith, integrating religious values into the core of their identity, with all other life needs viewed as subordinate to that ultimate concern. Koenig adapted this intrinsic framework for Items 3, 4, and 5, extracting core assertions from Hoge’s (1972) Intrinsic Religious Motivation Scale to construct a streamlined assessment of internalized faith orientation.

Simultaneously, the DUREL draws heavily upon the multi-factor sociological theories of religion pioneered by Charles Glock and Rodney Stark (1965). Glock and Stark established that religiosity consists of multidimensional components spanning ritualistic, ideological, intellectual, experiential, and consequential dimensions. Recognizing that public ritual behavior (ORA) and private ritual behavior (NORA) function along divergent psychosocial pathways, Koenig maintained their empirical autonomy rather than aggregating them into a single ritualistic score.

From a psychophysiological and behavioral perspective, the theoretical framework linking DUREL dimensions to health outcomes relies on several mechanisms:

• The Cognitive Appraisal and Coping Model: Derived from Lazarus and Folkman’s stress-coping paradigm, intrinsic religiosity provides a cognitive framework that facilitates positive cognitive reframing during catastrophic medical diagnoses, framing suffering within a meaningful existential narrative and reducing perceived threat and secondary neuroendocrine stress responses.
• The Social Integration and Capital Hypothesis: ORA operates predominantly through sociological networks, delivering practical tangible assistance, emotional validation, and structural accountability, which attenuate loneliness and dampen systemic hypothalamic-pituitary-adrenal (HPA) axis activation.
• Self-Regulation and Health Behavior Theory: High levels of intrinsic religiosity promote internal self-regulation and moral boundaries, leading to reduced consumption of harmful substances and improved adherence to medical protocols.

7. Validity

The psychometric validity of the DUREL has been extensively established across hundreds of empirical studies spanning psychiatric clinics, intensive care units, community cohorts, and cross-national populations.

Construct and Convergent Validity

Construct validity is evidenced by high correlations between the DUREL’s Intrinsic Religiosity subscale and established legacy instruments. Koenig and Büssing (2010) reported that the 3-item IR subscale correlates exceptionally well with Hoge’s 10-item Intrinsic Religious Motivation Scale (r = .85 to .92), indicating that the concise three-item formulation preserves the operational scope of the original construct. Furthermore, the DUREL subscales correlate positively with measures of general spirituality, such as the Spiritual Well-Being Scale (SWBS; r = .60 to .75) and the Santa Clara Strength of Religious Faith Questionnaire (r = .70 to .84).

Predictive and Criterion Validity

The predictive power of the DUREL has been demonstrated longitudinally within biomedical epidemiology. Numerous prospective cohort studies demonstrate that higher baseline scores on the ORA subscale (weekly or greater attendance) prospectively predict lower all-cause mortality over 10- to 20-year follow-up intervals, even after adjusting for baseline health status, socioeconomic status, and health-related behaviors (hazard ratios typically between 0.70 and 0.85). In clinical psychiatric populations, high intrinsic religiosity scores predict faster recovery from major depressive episodes among medically ill hospitalized older adults (Koenig et al., 1998). Conversely, individuals scoring low on the DUREL subscales frequently show heightened vulnerability to clinical anxiety, dysthymia, and lower subjective life satisfaction scores during acute physiological stressors.

Discriminant Validity

The necessity of treating the three DUREL dimensions as independent factors rather than an omnibus aggregate is justified by their divergent psychometric performance against physical health measures. For instance, in advanced geriatric and palliative oncology samples, NORA and IR often remain stable or increase as individuals utilize private prayer and theological reflection to navigate terminal illness. In contrast, ORA typically exhibits a steep decline due to physical functional impairment and mobility restrictions. If the five items were summed into a single score, physical disability would artifactually mask genuine levels of intrinsic religious commitment, validating the discriminant independence of these three subscales.

8. Reliability

The Duke University Religion Index exhibits excellent internal consistency and temporal stability across diverse populations, clinical contexts, and translated adaptations:

Internal Consistency

Because ORA (Item 1) and NORA (Item 2) are single-item behavioral metrics assessing specific behavioral frequencies, calculating internal consistency metrics across them is psychometrically inappropriate. However, the multi-item Intrinsic Religiosity (IR) subscale (Items 3, 4, and 5) consistently demonstrates robust internal consistency. In Dr. Koenig’s original clinical trials, Cronbach’s alpha for the IR subscale ranged from .75 to .85 in older medically ill cohorts. Subsequent large-scale epidemiological investigations have documented alpha coefficients ranging from .78 to .91.

Cross-cultural and linguistic validation efforts report comparable reliability metrics. The Persian translation of the DUREL evaluated by Saffari et al. (2013) among Muslim populations demonstrated an overall Cronbach’s alpha of .89, with the IR subscale demonstrating an alpha of .88. The validation of the Farsi version by Hafizi et al. (2013) confirmed an IR subscale alpha of .87. Similarly, Brazilian Portuguese validations in psychiatric and community samples (Luaccetti et al., 2011) yielded Cronbach’s alphas of .88 for the overall instrument and .86 for the intrinsic subscale. Spanish, German, and Chinese adaptations report similarly strong coefficients, typically falling between .78 and .90.

Test-Retest Temporal Stability

The temporal stability of the DUREL has been verified across varied retest intervals. A two-week test-retest evaluation among clinical outpatients and college cohorts yielded intra-class correlation coefficients (ICC) of .91 for ORA, .88 for NORA, and .92 for the IR subscale, indicating outstanding temporal reliability. Over longer observation periods (six months to one year), intrinsic religiosity demonstrates strong trait-like stability (r = .75 to .82), whereas ORA and NORA fluctuate in expected directions based on acute life crises, hospitalization, or environmental changes.

9. Factor Analysis

The latent dimensionality of the DUREL has been scrutinized via exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) across diverse international populations.

Exploratory Factor Analysis (EFA)

Initial exploratory factor analyses utilizing principal component and principal axis factoring with promax and varimax rotations confirmed that the five items load onto distinct factors that capture behavioral versus subjective-cognitive aspects of religiosity. Items 3, 4, and 5 consistently load heavily onto an Intrinsic Religiosity latent factor (factor loadings consistently exceeding .75 to .89), with minimal cross-loading onto items tracking institutional participation.

Confirmatory Factor Analysis (CFA) and Model Fit

Confirmatory factor analytic investigations have systematically tested alternative structural configurations: a single-factor unidimensional model, a two-factor model (aggregating ORA and NORA into a single “behavioral” factor alongside an “intrinsic” factor), and the theorized three-factor oblique model (ORA, NORA, and IR as three separate correlated dimensions). The three-factor model consistently shows superior statistical fit:

• Comparative Fit Index (CFI): Structural equation modeling studies (e.g., Saffari et al., 2013; Hafizi et al., 2013) report CFI values typically ranging from .97 to .99, substantially outperforming the single-factor model (which typically shows poor fit, with CFI < .85).
• Tucker-Lewis Index (TLI): TLI indices regularly exceed the .95 psychometric threshold, routinely reporting values between .96 and .99.
• Root Mean Square Error of Approximation (RMSEA): The three-factor specification generates RMSEA values between .03 and .06, indicating an excellent fit to the underlying empirical covariance structures.
• Standardized Factor Loadings: Within the three-factor model, standardized loadings for the Intrinsic Religiosity latent variable are exceptionally high: Item 3 typically loads between .76 and .84, Item 4 loads between .82 and .91, and Item 5 loads between .78 and .88.

Measurement invariance testing across sexes, age groups, and distinct religious traditions (e.g., comparing Christian, Jewish, and Muslim samples) indicates partial to strict metric and scalar invariance. This supports the validity of the DUREL’s factor structure across diverse research contexts.

10. Instrument / Measurement Tool

• Instrument Name: Duke University Religion Index (DUREL)
• Instrument Nature: Brief multidimensional self-report screening index
• Item Count: 5 items total
• Target Population: Adults, adolescents, medical inpatients, psychiatric cohorts, and general epidemiological survey respondents
• Estimated Completion Time: 1 to 2 minutes
• Subscale Architecture:
  • Subscale 1: Organizational Religious Activity (ORA): Item 1
  • Subscale 2: Non-Organizational Religious Activity (NORA): Item 2
  • Subscale 3: Intrinsic Religiosity (IR): Items 3, 4, and 5
• Response Formats:
  • Items 1 and 2: 6-point ordinal frequency scale:
    • 1 = Never (rarely or never)
    • 2 = Once a year or less (a few times a year)
    • 3 = A few times a year (once a week)
    • 4 = A few times a month (two or more times/week)
    • 5 = Once a week (daily)
    • 6 = More than once/week (more than once a day)
  • Items 3, 4, and 5: 5-point Likert agreement scale:
    • 1 = Definitely not true
    • 2 = Tends not to be true
    • 3 = Unsure
    • 4 = Tends to be true
    • 5 = Definitely true of me
• Scoring Methodology and Rules:
  • CRITICAL SCORING DIRECTIVE: The DUREL authors strongly discourage summing all five items into a single global score. Summing all items assumes a single-factor construct and obscures the distinct etiologic pathways through which communal activity (ORA), private practice (NORA), and intrinsic motivation (IR) relate to physical and mental health.
  • ORA Subscale Score: Derived directly from the score of Item 1 (Range: 1 to 6).
  • NORA Subscale Score: Derived directly from the score of Item 2 (Range: 1 to 6).
  • IR Subscale Score: Derived by summing the scores of Item 3, Item 4, and Item 5 (Range: 3 to 15).
  • Reverse Scoring: No reverse scoring is utilized; higher scores consistently reflect higher levels of religious involvement across all three subscales.

11. Permissions & Fee and Test Year

The Duke University Religion Index was developed in the late 1990s and published formally by Harold G. Koenig and colleagues in 1997, with comprehensive psychometric formulations consolidated in Hill and Hood’s Measures of Religiosity (1999) and later published in an open-access foundational review by Koenig and Büssing (2010).

Copyright and Licensing: The DUREL is placed in the public domain for clinical, educational, and academic research purposes. The scale may be utilized, reproduced, and translated without formal licensing fees, provided that appropriate academic citation and attribution are given to Dr. Harold G. Koenig and the Duke University Center for Spirituality, Theology and Health. Commercial software entities, pharmaceutical trials, and proprietary commercial test batteries should consult Duke University Medical Center or contact Dr. Koenig directly ([email protected]) to ensure adherence to institutional copyright guidelines.

12. References

Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and prejudice. Journal of Personality and Social Psychology, 5(4), 432–443. https://doi.org/10.1037/h0021212

Glock, C. Y., & Stark, R. (1965). Religion and society in tension. Rand McNally.

Hafizi, S., Memari, A. H., Pakrah, M., Mohebi, F., Saghazadeh, A., & Koenig, H. G. (2013). The Duke University Religion Index (DUREL): Validation and reliability of the Farsi version. Psychological Reports, 112(1), 151–159. https://doi.org/10.2466/08.PR0.112.1.151-159

Hill, P. C., & Hood, R. W., Jr. (Eds.). (1999). Measures of religiosity. Religious Education Press.

Hoge, D. R. (1972). A validated intrinsic religious motivation scale. Journal for the Scientific Study of Religion, 11(4), 369–376. https://doi.org/10.2307/1384677

Koenig, H. G., & Büssing, A. (2010). The Duke University Religion Index (DUREL): A five-item measure for use in epidemiological studies. Religions, 1(1), 78–85. https://doi.org/10.3390/rel1010078

Koenig, H. G., George, L. K., & Peterson, B. L. (1998). Religiosity and remission of depression in medically ill older patients. American Journal of Psychiatry, 155(4), 536–542. https://doi.org/10.1176/ajp.155.4.536

Lucchetti, G., Lucchetti, A. L. G., Peres, M. F., Leão, F. C., Moreira-Almeida, A., & Koenig, H. G. (2012). Validation of the Duke Religion Index: DUREL (Portuguese version). Journal of Religion and Health, 51(2), 579–586. https://doi.org/10.1007/s10943-010-9429-5

Saffari, M., Mohammadi Zeidi, I., Pakpour, A. H., & Koenig, H. G. (2013). Psychometric properties of the Persian version of the Duke University Religion Index (DUREL): A study on Muslims. Journal of Religion and Health, 52(2), 631–641. https://doi.org/10.1007/s10943-012-9636-9

13. Items of the Scale (Questionnaire)